PNN-based Rockburst Prediction Model and Its Applications

Abstract

Rock burst is one of main engineering geological problems greatly threatening the safety of construction. Prediction of rock burst is always an important issue concerning the safety of workers and equipments in tunnels. In this paper, a novel PNN-based rock burst prediction model is proposed to determine whether rock burst will happen in the underground rock projects and how much the intensity of rock burst is. The probabilistic neural network (PNN) is developed based on Bayesian criteria of multivariate pattern classification. Because PNN has the advantages of low training complexity, high stability, quick convergence, and simple construction, it can be well applied in the prediction of rock burst. Some main control factors, such as rocks' maximum tangential stress , rocks' uniaxial compressive strength , rocks' uniaxial tensile strength and elastic energy index of rock are chosen as the characteristic vector of PNN. PNN model is obtained through training data sets of rock burst samples which come from underground rock project in domestic and abroad. Other samples are tested with the model. The testing results agree with the practical records. At the same time, two realworld applications are used to verify the proposed method. The results of prediction are same as the results of existing methods, just same as what happened in the scene, which verifies the effectiveness and applicability of our proposed work. El fracturamiento o explosión de rocas es uno de los principales problemas en ingeniería geológica que amenaza significativamente la seguridad de una construcción. La predicción del fracturamiento de rocas es importante para la seguridad de los trabajadores y el equipamiento en túneles. En este artículo se propone un nuevo modelo de predicción de fracturamiento de rocas basado en una red neuronal probabilística (PNN por sus siglas en inglés) para determinar la posible ocurrencia e intensidad de uno de estos eventos en proyectos subterráneos. La PNN se desarrolló con base en un criterio Bayesiano para la clasificación multivariada de patrones. Debido a que la PNN tiene las ventajas de una menor complejidad de adiestramiento, estabilidad, rápida convergencia y simplicidad en su construcción, se puede adecuar en la predicción del fracturamiento de rocas. Algunos factores principales de control, como la fuerza máxima tangencial de rocas, la resistencia de compresión uniaxial, la fuerza de tensión uniaxial, y el índice de energía elástica de las rocas fueron escogidos como los vectores característicos de la PNN. El modelo se obtuvo a través del adiestramiento de datos sobre fracturamiento de rocas en proyectos subterráneos en diferentes localidades. Otras datos también se analizaron con el modelo. Los resultados de la evaluación se ajustan a los registros observados. Simultáneamente, se utilizaron dos aplicaciones prácticas para verificar el método propuesto. Los resultados de la predicción son similares a los de métodos existentes, un factor que además se presentó en las pruebas de campo, lo que demuestra la ef...

A Method for Rockburst Prediction in the Deep Tunnels of Hydropower Stations Based on the Monitored Microseismicity and an Optimized Probabilistic Neural Network Model

“…For example, neural networks, fuzzy comprehensive evaluation methods, analytic hierarchy processes, random forest methods, Bayesian models, support vector machines, cloud models, multivariate time series reconstruction, abstraction ant colony clustering algorithms, etc. [8,[27][28][29][30][31][32][33][34], have been used to study rock mechanics problems. Therefore, we can try to predict and can warn of rockburst risk by using the intelligent method and MS data together.…”

“…Feng et al [8] used MS information and a neural network to predict rockburst risk in deep tunnels of the Jinping II hydropower project. Zhou et al [34] successfully predicted the risk of rockburst in the construction of Tongyu and Qinling tunnels by using the probabilistic neural network (PNN) model; however, when using PNN for rockburst prediction, if the dimension of variables in input samples is large or there is a correlation between the variables, the prediction performance will be reduced [35], and the parameter smoothing factor in PNN affects the classification performance [36].…”

Prediction of Rockburst Intensity Grade in Deep Underground Excavation Using Adaptive Boosting Classifier

“…Five rockburst events in two different tunnel projects were predicted by the AdaBoost model. e field data were collected from available literature, including the Duoxiongla tunnel [58] and Anlu tunnel [59]. e prediction outcomes are summarized in Table 6, indicating that the rock burst intensity was predicted correctly for all cases.…”

A New Approach of Disaster Forecasting Based on Least Square Optimized Neural Network

“…The prediction results indicated that the PCA-GRNN model and PCA-RBF model showed more excellent network performances and higher prediction accuracy and generalization ability. Zhou and Wang [19] developed the probabilistic neural network (PNN) based on Bayesian criteria of multivariate pattern classification. Afraei et al [20] developed intelligent classification models for rock burst prediction by using five widespread techniques including artificial neural network techniques; different classification models are trained and tested with the same corresponding datasets to evaluate and compare their performances in the similar conditions.…”